Deterministic stress experiments

Know why your portfolio breaks, before it does.

PlanckNet decomposes portfolio losses into market moves, liquidity costs, and control failures. Same seed, same crash, every time. Fully debuggable.

Flight recorder for stress tests Decomposed PnL attribution Deterministic replay Kernel runs via REST API

Built for hedge funds and asset managers running multi-strategy portfolios where liquidity and controls dominate tail losses.

Round
Pre-Seed
Raising to deploy pilots and harden the platform.
Raise target
$200K on a SAFE
Detailed terms shared in investor materials.
Pilot pricing
$5K to $10K per month
90 days, hypothesis-driven deliverables.
Annual license
$100K to $150K ARR
Runs in your environment (SaaS, VPC, on-prem options).
Proof in 30 seconds

Determinism you can verify, not just claims.

Every run produces a structured artifact you can diff, replay, and audit. Determinism regression and verification checks keep the kernel honest.

Run artifact (field-level view)
Identical input yields identical output hash
Field Run A Run B
kernel_version v18.3.4.2 v18.3.4.2
seed 1337 1337
steps 20 20
run_hash 0x9c7a…f21b 0x9c7a…f21b

What you get per run

  • Decomposed PnL by step: market vs liquidity vs guard intervention
  • Replayable trace so failures are debuggable, not statistical
  • Verification checks to validate stability, drift, and regime integrity
Engineering appendix (spec excerpt)
spec excerpt (illustrative structure)
STRUCT EngineConfig:
  T_max: int
  W_drift_window: int
  B_CRIT_fin: float

FUNCTION LSO_apply_batched(X_block, prices, orders, depths):
  # clamp by depth, compute impact, compute holdings-based dPnL
  RETURN dPnL_LSO_block

FUNCTION enforce_stability_PER_LEN(...):
  # drift budgets per partition (LEN) with rolling windows
  RETURN drift_state, dPnL_stable

END-TO-END: determinism regression
  same seed -> identical PnL_logs, regime_logs, scenario_log
                
The problem

Most stress systems produce a report. They do not produce a replayable trace.

VaR aggregates risk. It does not explain it. When a strategy fails, you see the loss, but not the cause.

Black-box metrics

Aggregates hide failure modes. When something breaks, you cannot see what.

Liquidity blindness

Execution costs during stress are handled with crude add-ons, or ignored.

Non-reproducible tests

One-off stress reports cannot be replayed. You cannot debug randomness.

Core architecture

Decomposition is the product.

PlanckNet runs controlled experiments where each step separates market shock, liquidity cost, and control intervention.

RPO (Risk Propagation)

Models how factor shocks propagate through correlated instruments.

LSO (Liquidity Shock)

Quantifies execution cost and market impact under depth and urgency.

Drift Guard System

Testable circuit breakers with rolling PnL budgets per portfolio and partition.

Deterministic kernel

Seeded execution yields identical timelines for identical inputs. Replay and diff become first-class workflows.

Experimental layer

Not replacing vendor suites. PlanckNet runs alongside existing tools as the experimental, explainable stress layer.

Live demo

A guided, 2-minute proof.

Run the same scenario twice and confirm identical regime timeline and outputs.

PlanckNet demo
Hosted at fintech-8wji.onrender.com

VC preset (recommended)

  • Preset: Conservative
  • Seed: 1337
  • Steps: 20
  • What to look for: regime flip around steps 5 to 6, then decomposed PnL and guard status
  • Verification: run twice and confirm identical timeline
Ask questions after the demo
Note: The live demo is intended to show kernel behavior exposed via API and console output.
Business model

Paid pilots that test falsifiable hypotheses.

The goal is decision-making impact, not feature usage: reproducible stress reports with decomposition, plus an IC-ready walkthrough.

90-day pilot

$5K to $10K per month. Hypothesis-driven deliverables and reproducible artifacts.

Annual license

$100K to $150K ARR. Expansion path across desks, assets, and scenario packs.

Deployment

SaaS today, with VPC and on-prem options for institutional environments.

Initial buyers

Quant research, risk innovation teams, and advanced PMs with discretionary budgets and faster cycles.

Why now

Post-2023 governance pressure, liquidity-driven failures, rising strategy complexity, and brittle internal infra make reproducibility a requirement.

The ask

Raising to turn a validated kernel into institutional pilots.

This round funds validation, pilot deployments, and API hardening to reach 2 to 3 paid pilots and seed readiness.

Raise

$200K pre-seed on a SAFE. Details in investor materials.

Use of funds

Validation suite, pilots and conversion, infrastructure and legal, plus buffer.

9-month milestones

Validation and hardening, then 2 to 3 paid pilots, then first ARR and seed prep with case studies.

Request materials and book a call
Investor materials

Download and forward.

One-pager, pitch deck, and memo aligned to the same story as this site.

If you prefer a single email package, request it at office@coromandus.com.
Team

Research-grade systems that ship.

Three founders covering quantitative modeling, deterministic systems engineering, and institutional go-to-market. Full backgrounds shared on request.

Quant lead

Stress regime design, factor modeling, simulation logic, and methodology validation.

Engineering lead

Deterministic execution, artifact logging, API design, and verification discipline.

Product and BD

Pilot conversion, pricing, design partners, and institutional sales motion.

Contact

Book a call or request materials.

Email us and we will reply with scheduling options and the PDF package.

Email

office@coromandus.com
Include your firm, role, and availability.

Live demo

Go to demo section
Or open directly in a new tab from the demo card.